AI - harbinger of...????

[Nodvandigtid]

It's hurtling along at at real pace.

Here's an article that might of interest and of course another possible impact on us. 

 

AI-powered watermark removal poses uncomfortable implications for content use: Digital Photography Review (dpreview.com)

[Harry Harrison]

Just tried it on one of my own Alamy previews, a fairly messy scene, it removed most of the Alamy watermarks, just a few fragments remaining here and there. Another time when I wish this forum wasn't publicly viewable but I suppose the genie is out of the bottle. The comments under the Twitter post are interesting, but not necessarily in a good way.

[Harry Harrison]

I suppose Alamy saw this coming and so started to place more emphasis on chasing infringements.

[Allan Bell]

I have noted that the POD site I submit images to is accepting AI images.

 

Allan

 

[StokeCreative]

AI is evolving so rapidly, and as a design educator I feel the need to keep on top of it. Students will use it. As a content creator though its a bit scary.

I'm a fence sitter and indeed I have a few illustrations on Alamy created with MidJourney from my text prompts (but with keywords like 'controversy').

Photographers and Illustrators should perhaps be worried long term. The fact that these AI algorithms scrape billions of images means it would be very easy for any of our images to be 'used' without permission.

However at the end of the day, many new technologies have replaced or enhanced existing technology, workflow or indeed jobs. Photographers, artists, and musicians have always borrowed to some extent.  Some would say progress. I remain interested,' but sceptical.

 

Hugh

 

(P.S I'm not referring to removal of watermarks here, clearly wrong,  which I realise this post is primarily about)

 

[Jansos]

47 minutes ago, StokeCreative said:

AI is evolving so rapidly, and as a design educator I feel the need to keep on top of it. Students will use it. As a content creator though its a bit scary.

I'm a fence sitter and indeed I have a few illustrations on Alamy created with MidJourney from my text prompts (but with keywords like 'controversy').

Photographers and Illustrators should perhaps be worried long term. The fact that these AI algorithms scrape billions of images means it would be very easy for any of our images to be 'used' without permission.

However at the end of the day, many new technologies have replaced or enhanced existing technology, workflow or indeed jobs. Photographers, artists, and musicians have always borrowed to some extent.  Some would say progress. I remain interested,' but sceptical.

 

Hugh

 

(P.S I'm not referring to removal of watermarks here, clearly wrong,  which I realise this post is primarily about)

 

I feel that I should be interested, and to some extend I am, but it's beginning to feel as if I'm reaching one of those moments similar to that when you had to explain to one's elderly parents how to use the VHS player remote control/timer thingy! Have a go with this one... https://deepai.org/machine-learning-model/old-style-generator

Edited January 29 by Jansos
added URL

[Allan Bell]

10 minutes ago, Jansos said:

I feel that I should be interested, and to some extend I am, but it's beginning to feel as if I'm reaching one of those moments similar to that when you had to explain to one's elderly parents how to use the VHS player remote control/timer thingy!

 

Still can't get mine to work.

 

Allan

 

[M.Chapman]

Mmmm.... Beware???

According to "Captain Hook's" comments on the DPReview site, there's potentially a limited "rights grab" on any images submitted for processing

https://www.dpreview.com/news/0407669255/ai-powered-watermark-removal-poses-uncomfortable-implications-for-content-use

 

Reading the terms & conditions: (This company is based in India)

9. License and Use of your Content

9.1 You grant to us a royalty-free, worldwide, non-exclusive, sub-licensable and transferable license to use that Content including to reproduce, distribute, create derivative works, display, and perform it for any purpose in connection with the Service and our business, including for the purpose of promoting and redistributing part or all of the Service. This license is perpetual and authorizes us to make your Content available only for personal and non-commercial purposes, and only as permitted by the functionality of the Service.
"

I haven't download and installed the software, so I'm relying on the accuracy of Captain Hook's post. But, if you do download the software, check the licence terms!!

Given that the software claims to be "free and forever will be", it makes me wonder what's in it for the developers? An image rights grab?

 

Mark

[Harry Harrison]

47 minutes ago, M.Chapman said:

it makes me wonder what's in it for the developers?

Hopefully they won't find my rather unsuccessful offering any use but thanks for pointing that out. Their pricing, or rather I suppose their business model is weird, who is going to pay $299 a month to remove watermarks from up to 5000 images? Also strange that their 'Product of the day' No.2 rating is from March 2022.

[M.Chapman]

Oh - I missed their bulk processing options. 

12 hours ago, Harry Harrison said:

their business model is weird, who is going to pay $299 a month to remove watermarks from up to 5000 images?

Presumably folks who are up to no good and who are intent on depriving the image creators of their rights and income?

Should be illegal.

 

Mark

[Nodvandigtid]

‘ChatGPT said I did not exist’: how artists and writers are fighting back against AI | Artificial intelligence (AI) | The Guardian

[Harry Harrison]

Something from the BBC Sounds Tech Tent series, which I think is available worldwide. Contributions from Tim Flach, current president of AoP and renowned photographer of animals (5m 20secs). Followed by two representatives from SS AI.

 

https://www.bbc.co.uk/sounds/play/w3ct4kj9

 

 

Edited March 21 by Harry Harrison

[Nodvandigtid]

Photographer reveals prize-winning image was AI-made (rte.ie)

[Jill Morgan]

6 hours ago, Nodvandigtid said:

Photographer reveals prize-winning image was AI-made (rte.ie)

 

I am assuming there would be no EXIF data for the image.  This would have to be something that would now have to be checked before accepting submissions.  Stock agencies should need to add that to their algorithm before accepting images as photographs. 

 

Jill

[Jeffrey Isaac Greenberg]

> AI - harbinger of...????

 

AGI will probably not turn against man;

AGI will probably manipulate man &
turn man against man;

🤕WE    🤕FEEL IT   🤕HAPPENING ALREADY
👿 WHO LET THE AGI DOGS OUT?

[wiskerke]

15 minutes ago, Jeffrey Isaac Greenberg said:

> AI - harbinger of...????

 

AGI will probably not turn against man;

AGI will probably manipulate man &
turn man against man;

🤕WE    🤕FEEL IT   🤕HAPPENING ALREADY
👿 WHO LET THE AGI DOGS OUT?

Haha! Time to formulate the Three Laws of AI.

(Here's the Three Laws of Robotics as the model.)

Maybe ask ChatGPT to do it. 😁

 

wim

[John Mitchell]

For those of a certain age, can't say we weren't warned. 🙃

[wiskerke]

1 hour ago, John Mitchell said:

For those of a certain age, can't say we weren't warned. 🙃

🔴 I know I made some very poor decisions recently...

😁

 

wim

[John Mitchell]

1 hour ago, wiskerke said:

🔴 I know I made some very poor decisions recently...

😁

 

wim

 

Poor Hal. He didn't deserve to go that way. He had a heart after all. 😢

[gvallee]

Hubby asked AI to write an article about a certain subject. The results were impressive but hubby asked 'can you polish the article '. AI obliged and translated the whole article into Polish.

[Michael Ventura]

26 minutes ago, gvallee said:

Hubby asked AI to write an article about a certain subject. The results were impressive but hubby asked 'can you polish the article '. AI obliged and translated the whole article into Polish.

🤣😆😂

[Marianne]

21 hours ago, gvallee said:

Hubby asked AI to write an article about a certain subject. The results were impressive but hubby asked 'can you polish the article '. AI obliged and translated the whole article into Polish.

 

@Alamy BRING BACK THE LAUGHING EMOJI! PLEASE!

 

Edited May 13 by Marianne

[Marianne]

I've tried DALL-E, midjourney and Adobe Firefly. I like Firefly the best but it can't be used commercially and I don't like that they are using my stock images without compensation in the Beta version and who knows what pitance they will provide once it's out of Beta, but I still feel like my best course of action is to learn to use it and find a way to incorporate it into some of my work in the future.

 

In the meantime, I keep sending my daughter illustrations to share with my grandson. I did a series with pandas since he loves them and now he wants astronaut suits for his stuffed animals! 

 

it's fun to think up fantasy scenarios. The only time I tried a realistic photo (on Dall-E) it was surprisingly good and freaked me out. I'm still learning illustration, so using AI to help me achieve something I might not be able to do on my own is more satisfying. And that the pitfall here. It lets me do things I don't necessarily have the skills for on my own. But I also then do a lot of editing to make it my own. 

 

I've only submitted a couple of AI images (literally 2) that I made late last year to the other A when they first said they were accepting them. I prefer just playing with it for now and am having much better luck selling my own photos than any illustrations - AI assisted or not. I know there are contributors uploading thousands of new AI images that they've just created each week. The backlog there for illustrations is over a month, so I'm sticking with photos that pass inspection in 1-3 days or less.

 

I'm really glad Alamy isn't letting itself get overrun with AI right now. 

 

It's a new tool and I don't trust governments to protect our rights despite believing that wholesale image scraping without any form of compensation is wrong. I don't believe our copyright laws are set up to deal with this unique situation, which could not have been fathomed when such laws came into being and I just hope that Congress gets the right advice from all sides, particularly from those advocating for image, video and music creators to craft laws that protect our interests, but I'm not holding my breath. Unless it ends up creating problems for, say, the big movie studios, I'm not sure creators' rights will be adequately protected. Tech companies are running roughshod over our lives, gathering info that should appall our lawmakers, creating monopolies that insure small businesses need to follow their rules if they want their products to be seen, leaving empty storefronts everywhere, while those who should be looking out for the rest of us are in thrall to the tech companies and their money. 

 

So, how to respond? With resignation, most likely. So, I spend a few hours a week learning AI so when the time comes, I have some skills. But for now, it's all on my hard drive. 

[Jill Morgan]

17 hours ago, Marianne said:

I've tried DALL-E, midjourney and Adobe Firefly. I like Firefly the best but it can't be used commercially and I don't like that they are using my stock images without compensation in the Beta version and who knows what pitance they will provide once it's out of Beta, but I still feel like my best course of action is to learn to use it and find a way to incorporate it into some of my work in the future.

 

In the meantime, I keep sending my daughter illustrations to share with my grandson. I did a series with pandas since he loves them and now he wants astronaut suits for his stuffed animals! 

 

it's fun to think up fantasy scenarios. The only time I tried a realistic photo (on Dall-E) it was surprisingly good and freaked me out. I'm still learning illustration, so using AI to help me achieve something I might not be able to do on my own is more satisfying. And that the pitfall here. It lets me do things I don't necessarily have the skills for on my own. But I also then do a lot of editing to make it my own. 

 

I've only submitted a couple of AI images (literally 2) that I made late last year to the other A when they first said they were accepting them. I prefer just playing with it for now and am having much better luck selling my own photos than any illustrations - AI assisted or not. I know there are contributors uploading thousands of new AI images that they've just created each week. The backlog there for illustrations is over a month, so I'm sticking with photos that pass inspection in 1-3 days or less.

 

I'm really glad Alamy isn't letting itself get overrun with AI right now. 

 

It's a new tool and I don't trust governments to protect our rights despite believing that wholesale image scraping without any form of compensation is wrong. I don't believe our copyright laws are set up to deal with this unique situation, which could not have been fathomed when such laws came into being and I just hope that Congress gets the right advice from all sides, particularly from those advocating for image, video and music creators to craft laws that protect our interests, but I'm not holding my breath. Unless it ends up creating problems for, say, the big movie studios, I'm not sure creators' rights will be adequately protected. Tech companies are running roughshod over our lives, gathering info that should appall our lawmakers, creating monopolies that insure small businesses need to follow their rules if they want their products to be seen, leaving empty storefronts everywhere, while those who should be looking out for the rest of us are in thrall to the tech companies and their money. 

 

So, how to respond? With resignation, most likely. So, I spend a few hours a week learning AI so when the time comes, I have some skills. But for now, it's all on my hard drive. 

 

I have tried them all and still find Midjourney to give the best results for what I need.  So far Adobe Firefly has given me poor results for my requirements. I don't want photographic style images, I need cute funny illustrations that I put on products. And since the US will not give copyright on AI generated images, anyone can simply download them off of the AI site that is used.  Other governments are still to decide.  Here in Canada they are debating between the AI itself, the person who wrote the code, the person who created the prompt or maybe it is all public domain.  Still no decision.  I am thinking that overall it will end up that all AI generated images will be public domain.

 

They all still have problems doing animals.  If trying for a full body image, quite often they will give the animal (especially horses) extra legs and they will be bent in strange directions.  But I'm sure over time this will be no more.

 

There are people who sell their illustrations on various art sites and don't bother to even look closely at them before they submit.  I have seen them with Alamy logos as well as other logos from the watermarks on images that have been scraped. Signatures, etc.  Most images I have created still need work before I can use them.  The odd one comes out great without having to do much, but generally they all need tweeked as there will be flaws.

 

Jill